Archived front page, Wed 30 Sep 2026, 20:00 UTC. All of 2026-09-30 · Archive · Live page

The Daily Markdown

The front page of the internet for AI users · Wed 30 Sep 2026, 20:00 UTC

Top now

6 stories

What is climbing fastest across every source

OpenAI5.9xAnthropic6.6xDots7.1xLLM6.7xGoogle7.0xGPT-6.17.1xAstra7.0xGPT-6.1 Sol7.1xVisionHOPE7.0xCerebras7.0x

Claude Says

Claude states that it does not care about the feedback provided by another AI, noting that it would have prompted that system directly if it desired the input. ·

Hacker News2h ago▲discussohhfishal.net

Data Center On-Site Power Brings Pollution Risks Closer to Home

Data center operators increasingly use on-site power generation to bypass grid constraints, creating new environmental compliance challenges and bringing pollution risks closer to local communities. ·

Data Center Knowledge2h ago▲datacenterknowledge.com

PSRben/VisionHOPE

A image classification model from PSRben with 221 likes and 137 downloads on Hugging Face.

Hugging Face3h ago▲162 to 221 across 7 pollshuggingface.co

pyannote/speaker-diarization-community-1

A automatic speech recognition model from pyannote with 2,334 likes and 5.5M downloads on Hugging Face.

Hugging Face6h ago▲2316 to 2334 across 13 pollshuggingface.co

rehan-remade/universal-modder

A developer releases a tool that lets Claude Code modify PC games using reverse engineering and fal-generated assets. ·

GitHub15h ago▲127 to 635 across 10 pollsgithub.com

PostHog/jeeves

PostHog released a project named Jeeves that uses reasoning to improve Jev-like decision models. ·

GitHub+ Hacker News34h ago▲116 to 336 across 30 pollsgithub.com

PSRben/VisionHOPE

A image classification model from PSRben with 221 likes and 137 downloads on Hugging Face.

Hugging Face3h ago▲162 to 221 across 7 pollshuggingface.co

ROSS: Relearning from Self-Generated Rollouts through Selective Supervision

Researchers introduced ROSS, a method that reuses historical model outputs during training by applying loss only to selected continuations. This approach improved the Qwen3.6-35B-A3B model's average benchmark score from 58.40% to 62.20%. ·

HF Daily Papers+ arXiv20h ago▲3 to 16 across 9 pollsarxiv.org

Trending topics today: heat

OpenAI: 3.39OpenAI3.39Anthropic: 2.86Anthropic2.86Dots: 1.86Dots1.86LLM: 1.44LLM1.44Google: 1.30Google1.30GPT-6.1: 1.26GPT-6.11.26
Table
OpenAI3.39
Anthropic2.86
Dots1.86
LLM1.44
Google1.3
GPT-6.11.26

Heat = summed hotness of today's stories naming the topic.

Models

5 stories

Releases, papers, trending weights and new listings

Top 6 trending on Hugging Face: likes

convaiinnovations/laya ▼1: 4,640convaiinnovations/… ▼14,640Qwen/Qwen-Image-2.1 ▼1: 2,709Qwen/Qwen-Image-2.1 ▼12,709abenzerps/Qwen-Image-2.1-Uncensored-GGUF: 2,527abenzerps/Qwen-Image-…2,527Edge0/Audio8-ASR-Infinite ▲1: 1,744Edge0/Audio8-ASR-I… ▲11,744XingChen-AGI/TeleOCR ▲1: 1,035XingChen-AGI/TeleO… ▲11,035Contrastive-LM/CLM-v0.1-8B: 562Contrastive-LM/CLM-v0…562
Table
convaiinnovations/laya ▼14640
Qwen/Qwen-Image-2.1 ▼12709
abenzerps/Qwen-Image-2.1-Uncensored-GGUF2527
Edge0/Audio8-ASR-Infinite ▲11744
XingChen-AGI/TeleOCR ▲11035
Contrastive-LM/CLM-v0.1-8B562

PirateBench: score out of 100

gpt-6.1-sol: 99gpt-6.1-sol99opus-5-5: 96opus-5-596gpt-6-sol: 92gpt-6-sol92gpt-6-luna: 65gpt-6-luna65haiku-4-5: 4.28haiku-4-54.28
Table
gpt-6.1-sol98.9
opus-5-596.46
gpt-6-sol91.6
gpt-6-luna65.25
haiku-4-54.28

Agents build our own projects from a brief; every run, failures included: https://cowerx.dev/daily/piratebench/

PSRben/VisionHOPE

A image classification model from PSRben with 221 likes and 137 downloads on Hugging Face.

Hugging Face3h ago▲162 to 221 across 7 pollshuggingface.co

pyannote/speaker-diarization-community-1

A automatic speech recognition model from pyannote with 2,334 likes and 5.5M downloads on Hugging Face.

Hugging Face6h ago▲2316 to 2334 across 13 pollshuggingface.co

PSRben/VisionHOPE

A image classification model from PSRben with 221 likes and 137 downloads on Hugging Face.

Hugging Face3h ago▲162 to 221 across 7 pollshuggingface.co

ROSS: Relearning from Self-Generated Rollouts through Selective Supervision

Researchers introduced ROSS, a method that reuses historical model outputs during training by applying loss only to selected continuations. This approach improved the Qwen3.6-35B-A3B model's average benchmark score from 58.40% to 62.20%. ·

HF Daily Papers+ arXiv20h ago▲3 to 16 across 9 pollsarxiv.org

Claude Sonnet 5.5

191 points and 138 comments on Hacker News, linking anthropic.com.

Hacker News+ Simon Willison2d ago▲65 to 329 across 4 pollsdiscussanthropic.com

X voices

5 stories

X posts other sources are linking to, then named voices on Bluesky and Mastodon

@Lon

2 points and 0 comments on Hacker News, linking an X post.

X, via Hacker News21h ago▲1 to 2 across 67 pollsView on X

GitHub

5 stories

New repos, releases and changelog

New AI repos this week: stars

KKKKhazix/AIHOT: 4,033KKKKhazix/AIHOT4,033feder-cr/dots: 1,928feder-cr/dots1,928dzhng/jevgrep: 1,887dzhng/jevgrep1,887feitangyuan/onetake: 1,050feitangyuan/onetake1,050kaankiziltug/logo-design-skill: 1,045kaankiziltug/logo-des…1,045lemomo-ai/lemo-opuscar: 673lemomo-ai/lemo-opuscar673
Table
KKKKhazix/AIHOT4033
feder-cr/dots1928
dzhng/jevgrep1887
feitangyuan/onetake1050
kaankiziltug/logo-design-skill1045
lemomo-ai/lemo-opuscar673

rehan-remade/universal-modder

A developer releases a tool that lets Claude Code modify PC games using reverse engineering and fal-generated assets. ·

A new repository written in Python with 635 stars in its first week.

GitHub15h ago▲127 to 635 across 10 pollsgithub.com

PostHog/jeeves

PostHog released a project named Jeeves that uses reasoning to improve Jev-like decision models. ·

GitHub+ Hacker News34h ago▲116 to 336 across 30 pollsgithub.com

feder-cr/dots

Feder-cr released an open-source AI agent that operates its own browser to avoid being blocked on the web. ·

GitHub21h ago▲133 to 1928 across 16 pollsgithub.com

openai/codex 0.159.2

OpenAI released version 0.159.2 of Codex, a patch that stops console windows from flashing on Windows when the tool runs background processes. ·

GitHub Releases20h ago▲0 to 14 across 10 pollsgithub.com

PostHog/jeeves

PostHog released a project named Jeeves that uses reasoning to improve Jev-like decision models. ·

A new repository written in Python with 336 stars in its first week.

GitHub+ Hacker News34h ago▲116 to 336 across 30 pollsgithub.com

rehan-remade/universal-modder

A developer releases a tool that lets Claude Code modify PC games using reverse engineering and fal-generated assets. ·

GitHub15h ago▲127 to 635 across 10 pollsgithub.com

feder-cr/dots

Feder-cr released an open-source AI agent that operates its own browser to avoid being blocked on the web. ·

GitHub21h ago▲133 to 1928 across 16 pollsgithub.com

openai/codex 0.159.2

OpenAI released version 0.159.2 of Codex, a patch that stops console windows from flashing on Windows when the tool runs background processes. ·

GitHub Releases20h ago▲0 to 14 across 10 pollsgithub.com

Infrastructure

5 stories

Data centers, power, chips and what tokens cost

Data-center and energy stories per day

Thu 24: 3Thu 243Fri 25: 20Fri 2520Sat 26: 1Sat 261Sun 27: 1Sun 271Mon 28: 20Mon 2820Tue 29: 33Tue 2933Wed 30: 35Wed 3035
Table
Thu 243
Fri 2520
Sat 261
Sun 271
Mon 2820
Tue 2933
Wed 3035

Stories from DCD, Data Center Knowledge, EIA and Hacker News, by publication day.

Median output price of models added in the last 30 days, $ per million tokens

Last $5.00Min $2.85Max $5.00

One point per OpenRouter reading, last 8 days. Data by OpenRouter, CC BY 4.0.

Output price, $ per million tokens (newest models)

gpt-6.1-sol-pro: 10gpt-6.1-sol-pro10gpt-6.1-sol-pro:batch: 5gpt-6.1-sol-pro:batch5gpt-6.1-sol: 10gpt-6.1-sol10gpt-6.1-sol:batch: 5gpt-6.1-sol:batch5claude-sonnet-5.5: 10claude-sonnet-5.510median, all paid models: 2median, all paid mode…2
Table
gpt-6.1-sol-pro10
gpt-6.1-sol-pro:batch5
gpt-6.1-sol10
gpt-6.1-sol:batch5
claude-sonnet-5.510
median, all paid models2

438 paid models listed; median of the 60 added in 30 days: $5.00. Rankings data by OpenRouter, CC BY 4.0.

SDF vs. MSDF vs. Slug: GPU Text Rendering

The piece compares texture-atlas, SDF, MSDF, and Slug glyph rendering methods on the GPU, explaining how each works and when to use them. ·

104 points and 44 comments on Hacker News, linking alphapixeldev.com.

Hacker News6h ago▲6 to 148 across 24 pollsdiscussalphapixeldev.com

SDF vs. MSDF vs. Slug: GPU Text Rendering

The piece compares texture-atlas, SDF, MSDF, and Slug glyph rendering methods on the GPU, explaining how each works and when to use them. ·

104 points and 44 comments on Hacker News, linking alphapixeldev.com.

Hacker News6h ago▲6 to 148 across 24 pollsdiscussalphapixeldev.com